Ventilator treatment of the patient with acute respiratory failure (ARF) is aimed at opening up the lung, keeping it open, preserving gas exchange, and avoiding baro- and volotrauma (1–3). In the absence of a clinically acceptable method for determination of the functional residual capacity (FRC), surrogate indicators, such as the static pressure/volume curve and the upper and lower inflection points or the alveolar pressure/volume curve, have been used (4–6). Suter et al. (7) identified a positive end-expiratory pressure (PEEP) where maximum oxygen (O2) transport coincided with the highest static compliance and highest FRC. The FRC is thus a key for rational ventilator setting (8).
Measurements of FRC have hitherto mainly been obtainable in research situations either by dilution of gas with low solubility by rebreathing in a closed system, or by multiple breath washout technique. Suitable gases are sulfur hexafluoride (SF6), helium (He), O2, and nitrogen (N2). SF6, otherwise well suited, is not approved for human use by regulatory authorities (9,10). Helium is mainly used in the rebreathing technique, where the slow response of the detector is not a problem (11,12). O2 has the advantage of being easy to monitor with standard monitoring equipment, with a reasonable response time (13–15), but the O2 consumption must be corrected for, which is difficult at high inspiratory O2 fraction (Fio2). N2 can be measured with a mass spectrometer technique, but can also be calculated as the residual of O2 and carbon dioxide (CO2), which are the only other two components of ventilator gas (16–18). The main problem with both O2 and N2 as tracer gases is that changes of Fio2 of more than 0.2 fractions are required for adequate measurements of FRC (14,16). In severely ill ARF patients this has not been acceptable. Another problem inherent in the conventional N2 multiple breath washout (NMBW) technique using side stream gas analysis is the synchronization of gas flow and concentration measurements. This has to be performed prior to the continuous integration of flow and direct or indirect N2 concentration from O2 and CO2 measurements (17).
We have developed a NMBW using standard clinical O2 and CO2 sensors and flow meters to minimize the step change in O2 and washout period. The aim of this study was to validate the method for FRC measurements in an O2-consuming/CO2-producing lung model and to test the reproducibility in mechanically ventilated patients.
To avoid the synchronization problems we have focused on alveolar N2 exchange calculated from inspiratory and end-tidal plateau gas concentrations of O2 and CO2. A basic assumption is that inhomogeneity in alveolar gas distribution, reflected in steeply increasing or decreasing end-expiratory plateaus, is constant throughout the measurement procedure. Another assumption is that cellular metabolism and gas exchange between lung capillary blood and alveoli are stable during the washout/washin procedure.
O2 was analyzed using a side stream paramagnetic O2 analyzer with a response time of <480 ms (95% of full gain, manufacturer’s specification). CO2 was analyzed using a side stream infrared analyzer with a response time of <360 ms (95% of full gain, manufacturer’s specification). The gas analyzers were calibrated with a calibration gas with concentrations relevant to the measurements to be performed. After calibration the analyzers are automatically zeroed repeatedly to avoid baseline drift. Gas for breath-by-breath analysis of inspiratory and end-tidal concentrations was sampled at the y piece, and gas for analysis of mixed expired concentrations was sampled from a mixing box, 5 L volume, with a fan, connected to the expiratory outlet of the ventilator. Sampled gas was returned to the expiratory limb of the breathing system. Ventilation volumes were analyzed with a side stream spirometer based on an augmented Pitot methodology (19). The spirometer is zeroed automatically every second second. These analyzers were all part of an AS/3 modular monitoring system (Datex-Ohmeda, Helsinki, Finland).
Raw data of flow and gas concentrations were sampled with a frequency of 25 Hz and processed by the monitor software, which provides inspiratory and end-tidal gas concentrations, as well as tidal volume values, which then were collected by Collect S/5, ver. 4.0 software (Datex-Ohmeda) with a sampling frequency of 1 Hz. These data were exported to customized software, Testpoint (Capital Equipment Corp., Bedford, NH) for analysis and calculation of FRC.
Baseline O2 consumption (V̇o2) and CO2 production (V̇co2) were determined with indirect calorimetry measurements using the gas analyzer of the monitor in the mixing box for a period of 30 s in order to calculate the alveolar ventilation. The inspiratory minute ventilation (V̇i) was calculated by Haldane transformation from the measured expiratory minute ventilation (V̇e) with the assumption that there was no net exchange of N2. Expiratory alveolar minute ventilation, V̇ae was calculated according to Bohr’s formula:
The inspiratory alveolar minute ventilation (V̇ai) was calculated as the difference between V̇i and V̇e plus the expiratory alveolar minute ventilation:
The inspiratory and expiratory alveolar tidal volumes (Vtai and Vtae) were calculated from the alveolar minute ventilation and the breathing frequency. Breath-by-breath N2 exchange (VtN2) was calculated as the difference between inspired and expired N2 volume:
where FiN2 is the inspiratory N2 fraction, FetN2 is the end-tidal N2 fraction, Feto2 is the end-tidal O2 fraction, FiN2 = 1 − Fio2, and FetN2 = 1 − Feto2 − Fetco2
The inspiratory and end-tidal O2 and CO2 concentrations were acquired breath-by-breath from the monitor output.
where FiN2ini − FiN2end is the difference in inspiratory N2 concentration between start and end of washout. Calculations were based on a period of 3 time constants, when 95% of washin/washout was completed.
CO2 output was achieved by delivery of CO2 into the single alveolus of the model via a precision electronic flow controller. O2 consumption was achieved by combustion of hydrogen in a mini-Bunsen burner where 2H2 + O2 = 2H2O, i.e., the O2 consumption equals half of the delivered volume of hydrogen. An electronic flow regulator also controlled the hydrogen flow. Combustion took place in the single alveolus of the lung model (20,21). The respiratory quotient (RQ), which is the ratio V̇co2/V̇o2, could be varied freely by adjusting the settings of V̇co2 and V̇o2 of the lung model. The basal FRC of the lung model was 1.8 L and was increased stepwise to 2.8 and 3.8 L by addition of volume to the single alveolus.
With the lung model V̇co2/V̇o2 set at 200/200 and 200/240 mL/min, precision of FRC measurements was tested using step changes of Fio2 with 0.1, 0.2 or 0.3.
With FRC set at 1.8, 2.8 or 3.8 L the lung model was ventilated with an Fio2 of 0.4, 0.7, and 1.0. During measurements with lung model FRC set at 1.8 L, V̇co2/V̇o2 was set at 140/200, 170/200 and 140/140 mL/min, resulting in a RQ of 0.7, 0.85 and 1.0, respectively. During measurements with lung model FRC set at 2.8 L, V̇co2/V̇o2 was set at 200/280, 140/165 and 170/170 mL/min, resulting in a RQ of 0.7, 0.85 and 1.0, respectively. During measurements with lung model FRC set at 3.8 L, V̇co2/V̇o2 was set at 170/240, 200/235 and 200/200 mL/min, resulting in a RQ of 0.7, 0.85 and 1.0, respectively.
For FRC measurements, N2 washin was achieved by reducing Fio2 by 0.1 and N2 washout was achieved by changing Fio2 back to the original setting. Thus, each measurement was the average of a washin-washout procedure changing Fio2 sequentially from 0.4 to 0.3 to 0.4, from 0.7 to 0.6 to 07 or from 1.0 to 0.9 to 1.0.
The lung model was ventilated with a Siemens Servo 900C ventilator (Siemens-Elema, Solna, Sweden) at a respiratory rate of 12, 15 or 19/min and a minute ventilation set at 8, 10 or 12 L/min. During the evaluation of different step changes of Fio2, an Fio2 between 0.3 and 0.6 was used. In the evaluation of the effect of high Fio2 on FRC measurements, inspiratory O2 concentration was set at 0.4, 0.7, and 1.0.
The Ethical Committee of the Medical Faculty at Göteborg University approved the study and informed consent was obtained from the patients or next of kin. Measurements were performed in 28 patients whose demographics are presented in Table 1. All patients were ventilated with a Servo 900C or 300 ventilator in volume control mode, with Fio2 0.3–0.6, inspiration 25%, end-inspiratory pause 10% and a respiratory frequency of 12–20/min.
In 18 patients FRC was measured during a stable PEEP level already set for clinical reasons by changing Fio2 stepwise up and then back down by 0.1, 0.2 or 0.3 to achieve N2 washout/washin measurements. After a stabilization period the measurement was started, with a step increase in Fio2 of 0.3 to induce a washout of N2. After a new steady-state was reached, as indicated by the concentration difference between inspiratory and end-tidal O2 concentrations reaching the same level as before the start of the washout procedure, a step decrease of Fio2 of 0.3 to induce a washin of N2 was performed. This sequence was then repeated after steady-state was reached again, using a step change in Fio2 of 0.2, and then finally followed by a sequence using a step change in Fio2 of 0.1. The measurement sequence, using step changes of 0.3, 0.2 and 0.1, was then repeated. Each washout or washin procedure was analyzed during 3 time constants equivalent to 95% of a complete washin or washout procedure.
In 17 patients (7 of whom were among the 18 patients in whom different step changes of Fio2 were evaluated) the FRC was measured at 2 different PEEP levels, 5–8 cmH2O apart, by increasing and decreasing Fio2 by 0.1.
Results are presented as mean ± sd. Comparisons between patient measurements with different step changes of Fio2 as well as duplicate measurements were performed using Bland and Altman analysis (22).
The different FRCs of the lung model were measured with good precision using the modified NMBW method, irrespective of the level of Fio2 or ΔFio2 used (Figs. 1 and 2). When step changes of Fio2 of 0.1, 0.2 or 0.3 were used, minimal differences in precision were observed: 103 ± 5%, 101 ± 6% and 102 ± 4%, respectively, of the reference volume of the lung model. When ventilating the lung model with increasing Fio2 levels from 0.3, including levels of 0.7 and 1.0, where the O2 consumption was calculated from the CO2 production with a default RQ of 0.85, a very high precision of measurements was seen irrespective of the Fio2 used (Fig. 2). At Fio2s of 0.3–0.4, 0.7, and 1.0, the measured values were 100 ± 6%, 103 ± 8%, and 103 ± 7% of the reference FRC of the lung model. When the RQ of the lung model was varied between 0.7 and 1.0 using the default RQ value (0.85) for the NMBW algorithm, there was a small overestimation of FRC (116 ± 187 mL) at a true RQ of 0.7. At a true RQ of 0.85 the overestimation was only 36 ± 192 mL. In comparison, a true RQ of 1.0 resulted in a minimal underestimation of FRC (−19 ± 197 mL). These corresponded to 4%, 1.3% and −0.7% of the true FRC volume, respectively. The difference between washout and washin measurements in the lung model using a step change in Fio2 of 0.1 was 14 ± 187 mL, corresponding to 0.5% of the true FRC volume.
Analysis of each washin or washout procedure was performed during 3 time constants, which corresponded to a duration of about 80–120 s. In 2 patients with a very large FRC of more than 4 L the duration was >150 s. Comparisons in 28 patients of duplicate measurements of FRC (mean of a washin and washout procedure) at Fio2 steps of 0.1, 0.2, 0.25, and 0.3 showed a bias of −5 mL with a 95% confidence interval (CI) [-38, 29 mL] (Fig. 3). In 17 patients, measurements of FRC were performed as duplicate washin/washout procedures at 2 PEEP levels (∼ 7 cm H2O difference) using stepwise changes of Fio2 varying from 0.1 to 0.25. The bias of repeated measurements was −22 mL with a CI [-60, 16 mL] (Fig. 4). When comparing FRC measured using an Fio2 step change of 0.1 or 0.3 (mean of washin and washout), a bias of −9 mL with limits of agreement ± 356 mL was found (Fig. 5). Comparing the washin with the washout procedures using a step Fio2 of 0.1 resulted in a bias of 149 mL with limits of agreement of 484 mL.
We have modified and simplified a conventional NMBW method for measurement of functional residual capacity in ventilator-treated patients. The problems with continuous synchronization and integration of flow and gas concentration measurements have been circumvented by using the end-tidal and inspiratory O2 and CO2 concentration output signals from a standard side stream gas monitor and flow measurements from the ventilator. The method can be used with a step change of only 10% in inspiratory O2 concentration. In patients without chronic obstructive pulmonary disease (COPD), a washin or washout procedure takes less than 4 minutes to complete, resulting in a short and small change of alveolar O2 concentration. The method demonstrated high precision during lung model evaluation, and good reproducibility in patients; it is thus possible to use in patients with high inspiratory O2 concentrations as well.
The main problem concerning NMBW for routine clinical use is that N2 requires either a mass spectrometer or a Raman scattering technique for direct analysis. In patients in whom no gases other than O2, CO2, and N2 are present, N2 can be calculated as the residual by measurement of O2 and CO2. Fretschner et al. (16) presented a technique using a main-stream CO2 analyzer and a side-stream O2 analyzer. As the two wave forms are inversely congruent, they could be synchronized with a pneumotachograph measurement of gas flow for integration of flow and N2 concentrations. The synchronization procedure is very sensitive, however, especially when a substantial shift of inspiratory O2 concentration is made (30% in (16)). In this case, a small error in the synchronization may lead to gross miscalculations of N2 volume.
We used side-stream O2 and CO2 analyzers with marginally different response and delay times. By using only the plateau values of end-tidal and inspiratory gas concentrations, differences in delay and response times become unimportant, and there is no need for continuous synchronization. The correct calculation only requires that tidal flow/volume and inspiratory and end-tidal gas concentrations be tied to each other for each breath for correct calculations.
Eichler et al. (15) used a similar approach in a study of O2 washin/washout measurements of FRC. Their O2 sensor had a response time of 1.5 seconds, however, which is a limitation, as respiratory rate must be kept below 20/min for acceptable detection of end-tidal plateau values. Also, the slow response time of their O2 sensor affected the measurement of the first end-tidal values after changing O2 concentration, resulting in an overestimation of FRC of 400–500 mL.
The response time of <500 milliseconds of the O2 sensor in this study is fast enough to detect even the very first end-tidal plateau value correctly after making a step change in Fio2, and permits higher respiratory rates. The only necessary assumption is that the dead space for O2 and CO2 is equal (23). One could object that the difference in response time would result in the inspiratory and end-tidal O2 concentration being a little too small and large, respectively, and not comparable to the corresponding CO2 values. If a stepwise change in Fio2 causes a change in time constants of different parts of the lung, there could be an effect on the calculations of FRC. However, we saw no signs in the curve forms indicating such a time-constant change, and a reasonable assumption is that the lung compartment characteristics are identical before and after the washin or washout measurement procedure. Also, it must be noted that errors caused by differences in response time during a washin procedure will be counterbalanced by the same errors during the following washout procedure. The precision of our measurements is well within the limits proposed by Hedenstierna (8) specifically addressing FRC measurements (16).
To be able to measure FRC in critically ill patients with very high Fio2, the necessary step change in inspired O2 must be minimized. A 0.3 step change or more has been used in a number of studies (16–18). In the Eichler et al. (15) study of O2 washin/washout measurements of FRC, the method most similar to ours, a step change of 0.7 was used in mechanically ventilated patients. A change of such a magnitude in patients with 80–100% inspiratory O2 could severely affect O2 delivery and saturation and makes the method less suitable for use in the intensive care unit. We evaluated step changes in Fio2 of 0.1, 0.2, and 0.3 and found comparable results irrespective of size of change in Fio2 (Fig. 1). Even in the presence of severe ventilation/perfusion mismatch, the effect on arterial O2 content of a step change of Fio2 from 1.0 to 0.9 is likely tolerated in most cases. The measurement procedure, especially in acute respiratory distress syndrome (ARDS) patients with small FRC, is very short (<4 min). This is acceptable even in a critically ill patient. In the lung model, which is a single compartment model with good gas mixing, we chose to make the measurement over a three time constants period to cover 95% of the total washin/out. In ARDS patients with decreased FRC, this will be a fairly short period, but should be long enough to detect inhomogeneous lung pathology as well. The performance of the method in severely inhomogeneous lungs, i.e., COPD patients, was not evaluated in this study.
Our method is based on determination of baseline O2 consumption and CO2 production by indirect calorimetry, which is imprecise at Fio2>0.7 and impossible at an Fio2 of 1.0. Therefore we chose to calculate V̇co2 from mixed expiratory CO2 concentration and expiratory volume and to calculate V̇o2 from this V̇co2 with a default value of RQ of 0.85 when Fio2 was more than 0.7. For in vitro testing, our method and a similar method (18) require a lung model with O2 consumption and CO2 production. In the lung model evaluation we used V̇o2 and V̇co2 settings resulting in RQ of 0.7, 0.85 and 1.0. Measuring FRC with the default RQ of 0.85 as a base for calculations, when the model was ventilated with inspiratory O2 concentrations up to 100%, did not affect measurement precision, which indicates that RQ has a negligible effect on precision of FRC measurements. The metabolically active lung model used in our study (20,21) has the advantage of gases having the same humidity and temperature as airway gases of patients. The impact of humidity is not tested in more conventional lung model evaluations of FRC measurements (9,10,16–18). In this study we measured the baseline metabolic gas exchange in a mixing box on the expiratory outlet of the ventilator to make it possible for us to select different default values of RQ during the lung model study. However, in clinical practice, the capacity of the spirometry and gas module (MCOVX module, AS/3 and S/5, Datex-Ohmeda) to perform breath-by-breath indirect calorimetric gas exchange measurements can be used.
In spite of N2 having a very low solubility in blood and tissue, a certain amount of N2 diffuses from blood to the alveoli during a washout procedure and vice versa during a washin procedure, resulting in a slight overestimation of FRC measured by a single washout procedure and a slight underestimation of FRC measured by a single washin procedure. There is no consensus regarding the amount of tissue N2 diffusing in or out of the alveoli, but 40 mL/min has been proposed when changing the Fio2 by 0.8 (24). Our method only uses an Fio2 step change of 0.1, and the flux of tissue N2 will be low. We propose that a normal FRC measurement is found in a combined washout/washin procedure where the tissue N2 factor is eliminated, which is clearly demonstrated in the repeatability test with a very small bias when using a combined washin/washout procedure (Fig. 3). In a case where a single N2 washout measurement is performed, FRC will be overestimated with around 5%, and a single washin measurement will result in an underestimation of similar magnitude.
In conclusion, we have shown that FRC measurements with high precision can be obtained using a NMBW technique based on standard gas monitoring equipment and an Fio2 step change of as little as 0.1. The measurement technique has the potential of providing information on the lung volume status of severely ill ARDS patients ventilated with up to 100% O2 in a clinical setting as well.
The data collection program (Collect S/5) used in this study is commercially available, but the analysis software was specially customized for the study. At present, the analysis software needed to make automatic FRC measurements possible is under development (i.e., when Fio2 is increased or decreased stepwise, a measurement is automatically started), which will facilitate the clinical use.
1. Lachmann B. Open up the lung and keep the lung open. Intensive Care Med 1992;18:319–21.
2. Lichtwarck-Aschoff M, Mols G, Hedlund AJ, et al. Compliance is nonlinear over tidal volume irrespective of positive end-expiratory pressure level in surfactant-depleted piglets. Am J Respir Crit Care Med 2000;162:2125–33.
3. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000;342:1301–8.
4. Falke KJ, Pontoppidan H, Kumar A, et al. Ventilation with end-expiratory pressure in acute lung disease. J Clin Invest 1972;51:2315–23.
5. Sondergaard S, Karason S, Wiklund J, et al. Alveolar pressure monitoring: an evaluation in a lung model and in patients with acute lung injury. Intensive Care Med 2003;29:955–62.
6. Muramatsu K, Yukitake K. A new method to analyze lung compliance when pressure-volume relationship is nonlinear. Am J Respir Crit Care Med 1999;159:2028.
7. Suter PM, Fairley HB, Schlobohm RM. Shunt, lung volume and perfusion during short periods of ventilation with oxygen. Anesthesiology 1975;43:617–27.
8. Hedenstierna G. The recording of FRC: is it of importance and can it be made simple? Intensive Care Med 1993;19:365–6.
9. Jonmarker C, Jansson L, Jonson B, et al. Measurement of functional residual capacity by sulfur hexafluoride washout. Anesthesiology 1985;63:89–95.
10. Larsson A, Linnarsson D, Jonmarker C, et al. Measurement of lung volume by sulfur hexafluoride washout during spontaneous and controlled ventilation: further development of a method. Anesthesiology 1987;67:543–50.
11. Pelosi P, Cereda M, Foti G, et al. Alterations of lung and chest wall mechanics in patients with acute lung injury: effects of positive end-expiratory pressure. Am J Respir Crit Care Med 1995;152:531–7.
12. Patroniti N, Bellani G, Manfio A, et al. Lung volume in mechanically ventilated patients: measurement by simplified helium dilution compared to quantitative CT scan. Intensive Care Med 2004;30:282–9.
13. Stenqvist O, Olegard C, Sondergaard S, et al. Monitoring functional residual capacity (FRC) by quantifying oxygen/carbon dioxide fluxes during a short apnea. Acta Anaesthesiol Scand 2002;46:732–9.
14. Mitchell RR, Wilson RM, Holzapfel L, et al. Oxygen wash-in method for monitoring functional residual capacity. Crit Care Med 1982;10:529–33.
15. Eichler W, Schumacher J, Roth-Isigkeit A, et al. Automated evaluation of functional residual capacity by oxygen washout. J Clin Monit Comput 2002;17:195–201.
16. Fretschner R, Deusch H, Weitnauer A, Brunner JX. A simple method to estimate functional residual capacity in mechanically ventilated patients. Intensive Care Med 1993;19:372–6.
17. Wrigge H, Sydow M, Zinserling J, et al. Determination of functional residual capacity (FRC) by multibreath nitrogen washout in a lung model and in mechanically ventilated patients: accuracy depends on continuous dynamic compensation for changes of gas sampling delay time. Intensive Care Med 1998;24:487–93.
18. Zinserling J, Wrigge H, Varelmann D, et al. Measurement of functional residual capacity by nitrogen washout during partial ventilatory support. Intensive Care Med 2003;29:720–6.
19. Merilainen P, Hanninen H, Tuomaala L. A novel sensor for routine continuous spirometry of intubated patients. J Clin Monit 1993;9:374–80.
20. Stenqvist O, Sonander H, Lofstrom B, Nilsson K. An oxygen-consuming model lung for evaluation of anaesthetic circuits. Acta Anaesthesiol Scand 1982;26:322–6.
21. Svensson KL, Sonander HG, Stenqvist O. Validation of a system for measurement of metabolic gas exchange during anaesthesia with controlled ventilation in an oxygen consuming lung model. Br J Anaesth 1990;64:311–9.
22. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10.
23. Bartels J, Severinghaus JW, Forster RE, et al. The respiratory dead space measured by single breath analysis of oxygen, carbon dioxide, nitrogen or helium. J Clin Invest 1954;33:41–8.
24. Ruppel G. Manual of pulmonary function testing. 7th ed. St. Louis, MO.: Mosby, 1998.
© 2005 International Anesthesia Research Society
25. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138:720–3.